Time normalization of LPC feature using warping method

This paper presents pre-processing of input features to artificial neural network (NN). This is for preparation of reliable reference templates for the set of words to be recognized. The processed features are pitch and Linear Predictive Coefficients (LPC) for input and reference templates, based on...

Full description

Bibliographic Details
Main Authors: Sudirman, Rubita, Salleh, Sh. Hussain, Khalid, Puspa Inayat, Ahmad, Abd. Hamid
Format: Article
Language:English
Published: Faculty of Electrical Engineering, Universiti Teknologi Malaysia 2005
Subjects:
Online Access:http://eprints.utm.my/1308/
http://eprints.utm.my/1308/1/Elektrika05.pdf
Description
Summary:This paper presents pre-processing of input features to artificial neural network (NN). This is for preparation of reliable reference templates for the set of words to be recognized. The processed features are pitch and Linear Predictive Coefficients (LPC) for input and reference templates, based on Dynamic Time Warping (DTW) algorithm. The first task is to extract pitch features using Pitch Scale Harmonic Filter (PSHF) algorithm. Another task is to align the input frames (test set) to the reference template (training set) using DTW fixing frame (DTW-FF) algorithm. This proper time normalization is needed since NN is designed to compare data of the same length whilst same speech can varies in their length. By doing frame fixing (time normalization), the test set and the training set is adjusted to the same number of frames. Having both pitch and LPC features fixed frames, speech recognition using neural network can be performed.